Oecologia (2001) 127:87–96 DOI 10.1007/s004420000561 James T. Cronin · Warren G. Abrahamson Goldenrod stem galler preference and performance: effects of multiple herbivores and plant genotypes Received: 28 December 1999 / Accepted: 18 August 2000 / Published online: 24 November 2000 © Springer-Verlag 2000 Abstract Few studies have examined how the prefer- constraint on the discriminatory ability of female stem ence-performance relationship of an herbivore for differ- gallers preventing them from selecting the best hosts ent genotypes of its host plant is affected by the presence among plants that differ in genotype and level of envi- and/or feeding activity of other members of the herbi- ronmental stress (e.g., presence of interspecific herbi- vore assemblage. In an outdoor garden, we manipulated vores). the abundance of three common herbivores, the meadow spittlebug, a leaf beetle, and an aphid, on replicate 1-m2 Keywords Eurosta solidaginis · Gall insect · plots of 16 different genotypes of tall goldenrod, Solida- Herbivore assemblage · Host choice · go altissima. Adults of the goldenrod stem galler, Eur- Preference-performance relationship osta solidaginis, were subsequently released into the gar- den to oviposit among the host plants. Oviposition pref- erence was strongly influenced by plant genotype and Introduction the presence of two of the herbivores, spittlebugs and leaf beetles. The effects of the herbivores were additive: A simple expectation from evolutionary theory is that the presence of leaf beetles reduced preference by 6%, natural selection should favor an herbivore that preferen- spittlebugs by 18%, and both herbivores combined by tially oviposits on plant genotypes that yield high perfor- 25%. Plant genotype-herbivore species interaction ef- mance for their offspring (Thompson 1988). Positive fects on stem-galler preference, which would indicate the preference-performance correlations have been found in a presence of genetic variation among goldenrod geno- number of herbivorous insects (e.g., Craig et al. 1989; types in their norms of reaction for their acceptability as Price et al. 1990; Rossi and Strong 1991), but exceptions a host to the stem galler, were absent in this study. The are common (e.g., Karban and Courtney 1987; Courtney performance of the stem galler was also significantly af- and Kibota 1990; Fox 1993; Larsson et al. 1995). Factors fected by goldenrod genotype, but in general was not af- promulgated to explain these inconsistent results have run fected by the presence of herbivores early in the season the gamut from differences in plant apparency (Feeny (the exception was a positive correlation between the 1976; Chew and Courtney 1991), novel associations be- proportion of ramets infested by all herbivores and gall tween host plant and herbivore (Thompson 1988, 1996; size). Overall, we could find no correlation between Joshi and Thompson 1995), phenological differences in preference and performance. This is in accord with re- herbivore oviposition (Straw 1989; Briese 1996), varia- sults from previous studies on this system that were per- tions in herbivore abundance (Wiklund 1982), environ- formed in the absence of herbivores, suggesting that the mental predictability (Futuyma 1976; Cates 1981; Chew presence of herbivores in this study did not qualitatively and Courtney 1991; Lalonde and Roitberg 1992), and alter the preference-performance relationship. We sug- limited discriminatory ability of herbivores (host confu- gest that the lack of a positive correlation between host- sion hypothesis of Fox and Lalonde 1993; Larsson and plant preference and larval performance may reflect a Ekbom 1995). Regardless of the form, the preference- performance relationship can greatly influence the distri- J.T. Cronin (✉) bution and abundance of herbivore populations (e.g., Department of Biology, P.O. Box 9019, University of North Dakota, Grand Forks, ND 58202-9019, USA Price 1991, 1994; Ohgushi 1995; Bigger and Fox 1997), e-mail: [email protected] as well as the evolution of host-plant specificity, diet W.G. Abrahamson breadth, host-race formation, and sympatric speciation Department of Biology, Bucknell University, Lewisburg, (e.g., Bush 1975; Futuyma and Meyer 1980; Mitter et al. PA 17837, USA 1991; Joshi and Thompson 1995; Thompson 1996). 88 The presence of extrinsic factors that affect host-plant among early-season herbivore species. The effect of the preference and performance differentially may also alter herbivore assemblage on the preference-performance re- the preference-performance relationship. Water or nutri- lationship was also examined. Finally, we discuss how ents (Maddox and Cappuccino 1986; Horner and these herbivores may affect the distribution of, and what Abrahamson 1992, 1999; Preszler and Price 1995; role they may have played in the evolution of host-plant Ruohomaki et al. 1996), wind exposure (Cipollini 1997), choice by, the stem galler. amount of shade (Ruohomaki et al. 1996; Horner and Abrahamson 1992), presence of parasites and predators (Lawton and McNeill 1979; Strong and Larsson 1994; Materials and methods Stiling and Rossi 1996), or the presence of other herbi- vores (Lewis 1984; Faeth 1986; McMillin and Wagner Natural history of herbivores 1997; Cronin and Abrahamson 1999) may change through time or affect preference and performance in The life history, ecology and evolution of the goldenrod stem gall- er are detailed in Uhler (1951) and Abrahamson and Weis (1997); qualitatively different ways. One common situation only a brief description is provided here. In Pennsylvania, adult might involve the presence of an herbivore assemblage stem gallers oviposit into the terminal buds of the goldenrod, S. al- that feeds on the plant prior to the occurrence of the tar- tissima, around mid- to late May. Stem tissue shows signs of get herbivore. This assemblage might induce chemical swelling within 3 weeks and by mid-July the gall, harboring a sin- gle larva, reaches full size and is spheroid in shape. Larvae over- changes in the plant (e.g., Faeth 1986; Harrison and winter within the galls of senescent goldenrod ramets, then pupate Karban 1986; Karban and Adler 1996), alter plant ap- and eclose the following spring. In the absence of any other herbi- parency (e.g., by stunting growth), or directly deter the vores, stem gallers show strong differences in preference and per- colonization of the plant by the target herbivore, and so formance among goldenrod genotypes, but no positive correlation alter host-plant preference. Unless offspring performance between the two traits has been found (Anderson et al. 1989; Horner and Abrahamson 1992, 1999; Craig et al. 1999). is similarly affected, the preference-performance rela- The suite of herbivores that feed on goldenrod is extremely di- tionship may be fundamentally altered. For example, in a verse. According to Root and Cappuccino (1992), 138 species of greenhouse study (Cronin and Abrahamson 1999), we insects are capable of completing their development on S. altiss- found that meadow spittlebugs (Philaenus spumarius L.; ima. In addition to E. solidaginis, two species of leaf beetle (T. virgata and T. borealis), two aphids (U. nigrotuberculatum and Homoptera: Cercopidae), which feed on goldenrods ear- U. caligatum), and the meadow spittlebug (P. spumarius) com- ly in the season, drastically reduced host-plant prefer- prise the vast majority of the total herbivore biomass (Cappuccino ence by the goldenrod stem galler (Eurosta solidaginis 1987; Root and Cappuccino 1992; Meyer 1993). The spittlebug Fitch; Diptera: Tephritidae). In contrast, stem-galler sur- tends to be the first herbivore to begin feeding on goldenrods in the spring and all of these herbivores are present on goldenrods vivorship (a measure of performance) tended to increase before stem-galler adults begin to emerge (J.T. Cronin personal on spittlebug-infested plants. As a consequence of these observation). opposing trends, the relationship between preference and Several studies suggest that these herbivores can significantly performance went from no, to a slightly negative, corre- affect the quality and fitness of goldenrods, and consequently lation in the absence and presence of spittlebugs, respec- stem-galler preference and performance. Goldenrods that were ex- posed to feeding by either spittlebugs, leaf beetles or aphids for a tively. We are aware of no other studies that have ad- 3-week period had biomasses, specific leaf areas, growth rates, dressed this issue, particularly with regard to assemblag- photosynthetic rates, and seed production that were lower than es of herbivores that may act in concert to affect the plants that were free of herbivores (Meyer and Whitlow 1992; preference-performance relationship of another member Meyer 1993; Meyer and Root 1993). However, the effects of each herbivore on goldenrod ramets were not the same. For example, of that assemblage. spittlebugs, but not the other two herbivores, reduced the produc- In this study, we examined whether the assemblage of tion of lateral stems; and spittlebugs and leaf beetles, but not herbivores that feed on tall goldenrod (Solidago altiss- aphids, delayed flowering. In general, the magnitude of effects on ima) early in the season has an effect on the preference- the host plant were greatest for the spittlebugs, followed by leaf beetles; and to a much lesser extent, aphids. It remains untested performance relationship of the stem galler (E. solid- how each of these main herbivores of goldenrod